| 1. |
- Chatzimparmpas, Angelos, 1994-, et al.
(författare)
-
A survey of surveys on the use of visualization for interpreting machine learning models
- 2020
-
Ingår i: Information Visualization. - : Sage Publications. - 1473-8716 .- 1473-8724. ; 19:3, s. 207-233
-
Tidskriftsartikel (refereegranskat)abstract
- Research in machine learning has become very popular in recent years, with many types of models proposed to comprehend and predict patterns and trends in data originating from different domains. As these models get more and more complex, it also becomes harder for users to assess and trust their results, since their internal operations are mostly hidden in black boxes. The interpretation of machine learning models is currently a hot topic in the information visualization community, with results showing that insights from machine learning models can lead to better predictions and improve the trustworthiness of the results. Due to this, multiple (and extensive) survey articles have been published recently trying to summarize the high number of original research papers published on the topic. But there is not always a clear definition of what these surveys cover, what is the overlap between them, which types of machine learning models they deal with, or what exactly is the scenario that the readers will find in each of them. In this article, we present a metaanalysis (i.e. a ‘‘survey of surveys’’) of manually collected survey papers that refer to the visual interpretation of machine learning models, including the papers discussed in the selected surveys. The aim of our article is to serve both as a detailed summary and as a guide through this survey ecosystem by acquiring, cataloging, and presenting fundamental knowledge of the state of the art and research opportunities in the area. Our results confirm the increasing trend of interpreting machine learning with visualizations in the past years, and that visualization can assist in, for example, online training processes of deep learning models and enhancing trust into machine learning. However, the question of exactly how this assistance should take place is still considered as an open challenge of the visualization community.
|
|
| 2. |
- Chatzimparmpas, Angelos, 1994-, et al.
(författare)
-
VisRuler : Visual Analytics for Extracting Decision Rules from Bagged and Boosted Decision Trees
- 2023
-
Ingår i: Information Visualization. - : Sage Publications. - 1473-8716 .- 1473-8724. ; 22:2, s. 115-139
-
Tidskriftsartikel (refereegranskat)abstract
- Bagging and boosting are two popular ensemble methods in machine learning (ML) that produce many individual decision trees. Due to the inherent ensemble characteristic of these methods, they typically outperform single decision trees or other ML models in predictive performance. However, numerous decision paths are generated for each decision tree, increasing the overall complexity of the model and hindering its use in domains that require trustworthy and explainable decisions, such as finance, social care, and health care. Thus, the interpretability of bagging and boosting algorithms—such as random forest and adaptive boosting—reduces as the number of decisions rises. In this paper, we propose a visual analytics tool that aims to assist users in extracting decisions from such ML models via a thorough visual inspection workflow that includes selecting a set of robust and diverse models (originating from different ensemble learning algorithms), choosing important features according to their global contribution, and deciding which decisions are essential for global explanation (or locally, for specific cases). The outcome is a final decision based on the class agreement of several models and the explored manual decisions exported by users. We evaluated the applicability and effectiveness of VisRuler via a use case, a usage scenario, and a user study. The evaluation revealed that most users managed to successfully use our system to explore decision rules visually, performing the proposed tasks and answering the given questions in a satisfying way.
|
|
| 3. |
- Coimbra, Danilo B., et al.
(författare)
-
Explaining three-dimensional dimensionality reduction plots
- 2016
-
Ingår i: Information Visualization. - : Sage Publications. - 1473-8716 .- 1473-8724. ; 15:2, s. 154-172
-
Tidskriftsartikel (refereegranskat)abstract
- Understanding three-dimensional projections created by dimensionality reduction from high-variate datasets is very challenging. In particular, classical three-dimensional scatterplots used to display such projections do not explicitly show the relations between the projected points, the viewpoint used to visualize the projection, and the original data variables. To explore and explain such relations, we propose a set of interactive visualization techniques. First, we adapt and enhance biplots to show the data variables in the projected threedimensional space. Next, we use a set of interactive bar chart legends to show variables that are visible from a given viewpoint and also assist users to select an optimal viewpoint to examine a desired set of variables. Finally, we propose an interactive viewpoint legend that provides an overview of the information visible in a given three-dimensional projection from all possible viewpoints. Our techniques are simple to implement and can be applied to any dimensionality reduction technique. We demonstrate our techniques on the exploration of several real-world high-dimensional datasets.
|
|
| 4. |
- Einsfeld, Katja, et al.
(författare)
-
Knowledge Generation Through Human-Centered Information Visualization
- 2009
-
Ingår i: Information Visualization. - : Palgrave Macmillan Ltd. - 1473-8716 .- 1473-8724. ; 8:3, s. 180-196
-
Tidskriftsartikel (refereegranskat)abstract
- One important intention of human-centered information visualization is to represent huge amounts of abstract data in a visual representation that allows even users from foreign application domains to interact with the visualization, to understand the underlying data, and finally, to gain new, application-related knowledge. The visualization will help experts as well as non-experts to link previously or isolated knowledge-items in their mental map with new insights.Our approach explicitly supports the process of linking knowledge-items with three concepts. At first, the representation of data items in an ontology categorizes and relates them. Secondly, the use of various visualization techniques visually correlates isolated items by graph-structures, layout, attachment, integration, or hyperlink techniques. Thirdly, the intensive use of visual metaphors relates a known source domain to a less known target domain. In order to realize a scenario of these concepts, we developed a visual interface for non-experts to maintain complex wastewater treatment plants. This domain-specific application is used to give our concepts a meaningful background.
|
|
| 5. |
- Elmqvist, Niklas, 1977, et al.
(författare)
-
CiteWiz: a tool for the visualization of scientific citation networks
- 2007
-
Ingår i: Information Visualization. - 1473-8716 .- 1473-8724. ; 6:3, s. 215-232
-
Tidskriftsartikel (refereegranskat)abstract
- We present CiteWiz, an extensible framework for visualization of scientific citation networks. The system is based on a taxonomy of citation database usage for researchers, and provides a timeline visualization for overviews and an influence visualization for detailed views. The timeline displays the general chronology and importance of authors and articles in a citation database, whereas the influence visualization is implemented using the Growing Polygons technique, suitably modified to the context of browsing citation data. Using the latter technique, hierarchies of articles with potentially very long citation chains can be graphically represented. The visualization is augmented with mechanisms for parent–child visualization and suitable interaction techniques for interacting with the view hierarchy and the individual articles in the dataset. We also provide an interactive concept map for keywords and co-authorship using a basic force-directed graph layout scheme. A formal user study indicates that CiteWiz is significantly more efficient than traditional database interfaces for high-level analysis tasks relating to influence and overviews, and equally efficient for low-level tasks such as finding a paper and correlating bibliographical data.
|
|
| 6. |
- Elmqvist, Niklas, 1977, et al.
(författare)
-
DataMeadow: a visual canvas for analysis of large-scale multivariate data
- 2008
-
Ingår i: Information Visualization. - : SAGE Publications. - 1473-8716 .- 1473-8724. ; 7:1, s. 18-33
-
Tidskriftsartikel (refereegranskat)abstract
- Supporting visual analytics of multiple large-scale multidimensional data sets requires a high degree of interactivity and user control beyond the conventional challenges of visualizing such data sets. We present the DataMeadow, a visual canvas providing rich interaction for constructing visual queries using graphical set representations called DataRoses. A DataRose is essentially a starplot of selected columns in a data set displayed as multivariate visualizations with dynamic query sliders integrated into each axis. The purpose of the DataMeadow is to allow users to create advanced visual queries by iteratively selecting and filtering into the multidimensional data. Furthermore, the canvas provides a clear history of the analysis that can be annotated to facilitate dissemination of analytical results to stakeholders. A powerful direct manipulation interface allows for selection, filtering, and creation of sets, subsets, and data dependencies. We have evaluated our system using a qualitative expert review involving two visualization researchers. Results from this review are favorable for the new method.
|
|
| 7. |
- Elmqvist, Niklas, et al.
(författare)
-
Fluid Interaction for Information Visualization
- 2011
-
Ingår i: Information Visualization. - : SAGE Publications. - 1473-8716 .- 1473-8724. ; 10:4, s. 327-340
-
Tidskriftsartikel (refereegranskat)abstract
- Despite typically receiving little emphasis in visualization research, interaction in visualization is the catalyst for the user’s dialogue with the data, and, ultimately, the user’s actual understanding and insight into these data. There are many possible reasons for this skewed balance between the visual and interactive aspects of a visualization. One reason is that interaction is an intangible concept that is difficult to design, quantify, and evaluate. Unlike for visual design, there are few examples that show visualization practitioners and researchers how to design the interaction for a new visualization in the best manner. In this article, we attempt to address this issue by collecting examples of visualizations with ‘best-in-class’ interaction and using them to extract practical design guidelines for future designers and researchers. We call this concept fluid interaction, and we propose an operational definition in terms of the direct manipulation and embodied interaction paradigms, the psychological concept of ‘flow’, and Norman’s gulfs of execution and evaluation.
|
|
| 8. |
- Forsell, C., et al.
(författare)
-
Surface Glyphs for Efficient Visualization of Multivariate Data
- 2006
-
Ingår i: Information Visualization. - 1473-8716 .- 1473-8724. ; 5:2, s. 112-124
-
Tidskriftsartikel (refereegranskat)abstract
- We present a first effort to evaluate the possible utility of a new type of surface glyphs intended for visualizations of multivariate spatial data. The glyphs are based on results from vision research suggesting that our perception of metric 3D structure is distorted and imprecise relative to the actual scene before us; only a class of qualitative properties of the scene is perceived with accuracy. These properties are best characterized as being invariant over affine but not Euclidean transformations. A large number of possible 3D glyphs for the visualization of spatial data can be constructed using such properties. One group of such glyphs is based on the local sign of surface curvature. We investigated this group in two visualization experiments. The results show that available sources of 3D structural information were sufficient for our subjects to make fast and accurate judgments. Some implications for visualization are also discussed.
|
|
| 9. |
- Forsén, Gunnar, et al.
(författare)
-
Pinpoint : A design study in interactive visualization for finding people in a large organization
- 2009
-
Ingår i: Information Visualization. - : Palgrave Macmillan. - 1473-8716 .- 1473-8724. ; 9, s. 141-151
-
Tidskriftsartikel (refereegranskat)abstract
- This design study reports on how information visualization was applied to the common problem for knowledge workers to find people with relevant expertise and interests in a large organization. The outcome is the Pinpoint concept, an interactive visualization where the user's most closely related colleagues are presented radially for browsing, filtering and further exploration of the topical networks of the organization. Based on design reasoning and empirical validation, we claim that the concept approaches desirability as well as usefulness and usability in its intended use situation. We further argue that it is feasible for deployment under certain conditions, and that it is applicable in a range of large organizations if organization-specific policies and standards are taken into account.
|
|
| 10. |
- Gleicher, Michael, et al.
(författare)
-
Visual Comparison for Information Visualization
- 2011
-
Ingår i: Information Visualization. - : SAGE Publications. - 1473-8716 .- 1473-8724. ; 10:5-4, s. 289-309
-
Tidskriftsartikel (refereegranskat)abstract
- Data analysis often involves the comparison of complex objects. With the ever increasing amounts and complexity of data, the demand for systems to help with these comparisons is also growing. Increasingly, information visualization tools support such comparisons explicitly, beyond simply allowing a viewer to examine each object individually. In this paper, we argue that the design of information visualizations of complex objects can, and should, be studied in general, that is independently of what those objects are. As a first step in developing this general understanding of comparison, we propose a general taxonomy of visual designs for comparison that groups designs into three basic categories, which can be combined. To clarify the taxonomy and validate its completeness, we provide a survey of work in information visualization related to comparison. Although we find a great diversity of systems and approaches, we see that all designs are assembled from the building blocks of juxtaposition, superposition and explicit encodings. This initial exploration shows the power of our model, and suggests future challenges in developing a general understanding of comparative visualization and facilitating the development of more comparative visualization tools.
|
|
